Borgogno, Luca (2024) Word Surprisal effect on Cognitive Processes During Reading: An EEG and Eye-Tracking Study. Bachelor's Thesis, Artificial Intelligence.
|
Text
finalthesis.pdf Download (2MB) | Preview |
|
![[img]](https://fse.studenttheses.ub.rug.nl/style/images/fileicons/text.png) |
Text
toestemming_ Luca Borgogno _ degree programme_ Artificial Intelligence.pdf Restricted to Registered users only Download (128kB) |
Abstract
This paper investigates the brain’s prediction mechanisms in natural language pro- cessing by examining how word surprisal affects cognitive processes during naturalistic reading. Using EEG and eye-tracking technologies, this research measures brain activity and pupil di- lation while participants read sentences from ”Animal Farm” by George Orwell. This study quantified word surprisal using the GPT-2 language model, aiming to simulate the brain’s pre- dictive mechanisms. The EEG analysis focuses on the N400 and P600 components, associated with semantic and syntactic processing respectively. However, the results did not show significant differences in neural responses across varying levels of word surprisal. Pupil dilation data sug- gested a correlation between higher surprisal and increased cognitive load, although these results were not conclusive due to anomalies in the data indicating the need for further investigation. This study highlights the complexity of modelling the brain’s predictive mechanisms and sug- gests improvements for future research to better understand how word predictability influences language comprehension during natural reading.
| Item Type: | Thesis (Bachelor's Thesis) |
|---|---|
| Supervisor name: | Borst, J.P. and Krause, J. |
| Degree programme: | Artificial Intelligence |
| Thesis type: | Bachelor's Thesis |
| Language: | English |
| Date Deposited: | 16 Jul 2024 14:04 |
| Last Modified: | 16 Jul 2024 14:04 |
| URI: | https://fse.studenttheses.ub.rug.nl/id/eprint/33472 |
Actions (login required)
 |
View Item |